This paper investigates the macroeconomic effects of fiscal policy in a setting in which private agents receive noisy signals about future shocks to government expenditures. We show how to empirically identify the relative weight of news and noise shocks to government spending and compute the level of noise for Canada, the UK and the US.We then investigate the quantitative implications of imperfect fiscal policy information using a medium-scale DSGE model. We find that when the government seeks to implement a persistent change in expected public spending, the existence of noise (as estimated using actual data) implies a sizable difference in fiscal multipliers compared to the perfect fiscal foresight case.
Government spending; Noisy Information; DSGE Models;
TSE Working Paper, n° 16-634, mars 2016